Discrimination surfaces with application to region-specific brain asymmetry analysis

Area of Science:

• Neuroimaging
• Biostatistics
• Psychiatry

Background:

• Identifying brain regions associated with diseases like schizophrenia is crucial for diagnosis.
• Current methods may lack precision in localizing these discriminative areas.
• Understanding brain asymmetry is key in neurological and psychiatric research.

Purpose of the Study:

• Introduce discrimination surfaces as a novel diagnostic tool.
• Localize brain regions with higher discrimination between diseased and non-diseased individuals.
• Investigate differences in local asymmetry scores in schizophrenic patients.

Main Methods:

• Developed a Mann-Whitney type statistic for random fields.
• Applied the statistic to estimate discrimination surfaces.
• Analyzed asymptotic behavior and simulation results for accuracy.
• Conducted empirical analysis on schizophrenic patients and control groups.

Main Results:

• The proposed estimator accurately recovers the true discrimination surface.
• Identified intervals of maximal discrimination within the brain.
• Schizophrenic patients showed lower local asymmetry scores in anterior brain regions compared to controls.

Conclusions:

• Discrimination surfaces offer a powerful tool for localizing disease-specific brain regions.
• The findings suggest altered local asymmetry in the anterior brain in schizophrenia.
• This approach has potential for improving diagnostic tools in psychiatry.